Datasets:
Formats:
csv
Size:
1K - 10K
Tags:
tabular
long-context-language-modeling
multi-head-attention
knowledge-graphs
neuro-symbolic-learning
causal-intervention
License:
| mode,unit_of_assignment,historical_state,admission_rule,exported_computation,stores_global_token_kv,requires_typed_graph_program | |
| GLOBAL,Global KV head,Full causal token KV history,"Retain when late-context lexical or discourse utility is material, graph substitution is inadequate, or cluster-retention constraints require the head.",Compact global multi-head-attention projection with historical KV,true,false | |
| LOCAL,Global KV head before export; local head after export,1024-token local-window KV,"The head passes the pre-specified candidate rule, is selected jointly under the interaction and cluster constraints, and has no event-scoped graph program.",Local causal attention with no full-history token KV,false,false | |
| LOCAL_GRAPH,Global KV head before export; local graph head after export,1024-token local-window KV plus an event-scoped typed result,"The head passes the pre-specified candidate rule, is selected jointly, has a valid typed program assignment, and shows relation-specific causal handoff without wrong-event or random benefit.",Local causal attention plus exact event-scoped typed graph read with no full-history token KV,false,true | |